Skipping cluster bootstrapping as local node does not match bootstrap requirements – How to solve related issues

Overview

Elasticsearch has many settings that can cause significant performance problems if not set correctly. To prevent this happening, Elasticsearch carries out “bootstrap checks” to ensure that these important settings have been covered. If any of the checks fail, Elasticsearch will write an error to the logs and will not start. In this guide we cover common bootstrap checks you should know and how to configure your settings correctly to pass the checks successfully.

Bootstrap checks are carried out when the network.host setting in:

network.host: 0.0.0.0

If network host is not set and you use the default localhost, then Elasticsearch will consider the node to be in development mode, and bootstrap checks will not be enforced.

Common issues with bootstrap checks

If you install elasticsearch using RPM or Debian packages (strongly recommended) then most of the important configuration is already done for you, and the only bootstrap checks you are likely to run up against are the following.

Heap size check

The minimum and maximum heap sizes specified in jvm.options (or via environment variables) must be equal to one another.

File descriptors check

Minimum file descriptors must have been set to at least 65535

Memory lock check

There are various methods used to prevent memory swapping, and you must use one of them.  The most common is to set in elasticsearch.yml

bootstrap.memory_lock: true

For this to be effective you must give permission to elasticsearch to enable this. There are various ways to do so, depending upon your operating system.

Discovery configuration checks

At least one of the following properties must be configured in elasticsearch.yml to ensure that your node can form a cluster properly:

• discovery.seed_hosts
• discovery.seed_providers
• cluster.initial_master_nodes

Less common bootstrap check issues

If you are not using RPM or debian packages, you may come across the following issues:

Max number of threads check

You must allow your system to create at least 4096 threads. In linux this is done by editing /etc/security/limits.conf and adjusting the nproc setting.

Max file size check

Your system should be able to create unlimited file sizes. In linux this is done by editing /etc/security/limits.conf and adjusting the fsize setting

Max virtual memory check

The system should be able to create unlimited virtual memory for the elasticsearch user.

This is done by editing /etc/security/limits.conf 

<user> - as unlimited

Max map count check

The system must be able to use mmap effectively.  This is done by running the command 

sysctl vm.max_map_count 262144

Other checks

The following checks are also carried out, but are rarely found:

OsX File Descriptor Check
Client Jvm Check
UseS erial GC Check
System Call Filter Check
Might Fork Check
On Error Check
On Out Of Memory Error Check
Early Access Check
G1GC Check
All Permission Check

Overview

An Elasticsearch cluster consists of a number of servers (nodes) working together as one. Clustering is a technology which enables Elasticsearch to scale up to hundreds of nodes that together are able to store many terabytes of data and respond coherently to large numbers of requests at the same time.

Search or indexing requests will usually be load-balanced across the Elasticsearch data nodes, and the node that receives the request will relay requests to other nodes as necessary and coordinate the response back to the user.

Notes and good things to know

The key elements to clustering are:

Cluster State – Refers to information about which indices are in the cluster, their data mappings and other information that must be shared between all the nodes to ensure that all operations across the cluster are coherent.

Master Node – Each cluster must elect a single master node responsible for coordinating the cluster and ensuring that each node contains an up-to-date copy of the cluster state.

Cluster Formation – Elasticsearch requires a set of configurations to determine how the cluster is formed, which nodes can join the cluster, and how the nodes collectively elect a master node responsible for controlling the cluster state. These configurations are usually held in the elasticsearch.yml config file, environment variables on the node, or within the cluster state.

Node Roles – In small clusters it is common for all nodes to fill all roles; all nodes can store data, become master nodes or process ingestion pipelines. However as the cluster grows, it is common to allocate specific roles to specific nodes in order to simplify configuration and to make operation more efficient. In particular, it is common to define a limited number of dedicated master nodes.

Replication – Data may be replicated across a number of data nodes. This means that if one node goes down, data is not lost. It also means that a search request can be dealt with by more than one node.

Common problems

Many Elasticsearch problems are caused by operations which place an excessive burden on the cluster because they require an excessive amount of information to be held and transmitted between the nodes as part of the cluster state. For example:

• Shards too small
• Too many fields (field explosion)

Problems may also be caused by inadequate configurations causing situations where the Elasticsearch cluster is unable to safely elect a Master node. This situation is discussed further in: 

• Master node not discovered
• Split brain problem

Backups

Because Elasticsearch is a clustered technology, it is not sufficient to have backups of each node’s data directory. This is because the backups will have been made at different times and so there may not be complete coherency between them. As such, the only way to backup an Elasticsearch cluster is through the use of snapshots, which contain the full picture of an index at any one time.

Cluster resilience

When designing an Elasticsearch cluster, it is important to think about cluster resilience. In particular – what happens when a single node goes down? And for larger clusters where several nodes may share common services such as a network or power supply – what happens if that network or power supply goes down? This is where it is useful to ensure that the master eligible nodes are spread across availability zones, and to use shard allocation awareness to ensure that shards are spread across different racks or availability zones in your data center.

Overview

To put it simply, a node is a single server that is part of a cluster. Each node is assigned one or more roles, which describe the node’s responsibility and operations. Data nodes store the data, and participate in the cluster’s indexing and search capabilities, while master nodes are responsible for managing the cluster’s activities and storing the cluster state, including the metadata.

While it is possible to run several node instances of Elasticsearch on the same hardware, it’s considered a best practice to limit a server to a single running instance of Elasticsearch.

Nodes connect to each other and form a cluster by using a discovery method. 

Roles

Master node

Master nodes are in charge of cluster-wide settings and changes – deleting or creating indices and fields, adding or removing nodes and allocating shards to nodes. Each cluster has a single master node that is elected from the master eligible nodes using a distributed consensus algorithm and is reelected if the current master node fails.

Coordinating (client) node

There is some confusion in the use of coordinating node terminology. Client nodes were removed from Elasticsearch after version 2.4 and became coordinating nodes.

Coordinating nodes are nodes that do not hold any configured role. They don’t hold data and are not part of the master eligible group nor execute ingest pipelines. Coordinating nodes serve incoming search requests and act as the query coordinator running query and fetch phases, sending requests to every node that holds a shard being queried. The coordinating node also distributes bulk indexing operations and route queries to shards based on the node’s responsiveness.